May I know what does repeated significance testing means in TSA manual?
For binary outcome (mortality), we only insert n/N from each trial into the revMan to meta-analytic the result. If there are some new trials recently, we will added n/N of new trials into the existing table? However, I am not understand how this update meta-analysis will contribute to problem of repeated significance testing in TSA as we will re-do statistical test when new data are added?
Kindly please advise me.
The point of sequential analysis is that the sample size is not fixed and you collect data until significant results are observed. This would involve significance testing every time you collected a piece of data.
I hope this answered your question.
Hi Ben,
Thank you very much for your answer.
Don't we re-analysed all the data when we update a meta analysis after any recent trials have been published? I can't figure out how is this updated meta-analysis contribute to the risk of random error due to repeated significance testing as listed in the user manual. I thought we will perform only one significance testing on updated meta analysis?
Many thanks for your attention and I look forward to your guidance.
Have a good weekend.
Ka
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis